Image compression using adaptive wavelet packet for zero tree coding and SPIHT

Abstract

Image coding methods based on adaptive wavelet transform and those employing zerotree quantization have been shown to be successful in recent years. Wavelet based zerotree image coding methods exploit the similarities across wavelet subbands by grouping coefficients belonging to subbands of different resolutions and encoding the group as a single codeword. The zerotree coding schemes are quite efficient in terms of both computational complexity and compression performance. Wavelet based methods have a problem at low bit rates. Wavelet packets pinpoint signal components present locally in the frequency domain. The nondyadic nature of a wavelet packet (WP) transform allows us to find an orthogonal basis adapted to the contents of given image and to the purpose of representation In this paper, we present a general zerotree structure for an arbitrary wavelet packet geometry in an image coding framework. A fast basis selection algorithm which uses a Markov chain based cost estimate of encoding the image using this structure is developed.We compare the AWPZT coding with SPIHT coding method. As a result, our adaptive wavelet zerotree image coder has a relatively low computational complexity, performs comparably to the state- of-the-art image coders in terms of visual quality at low bit rates, overcomes parenting conflict and is capable of progressively encoding images.